计算机工程

• 开发研究与工程应用 • 上一篇    下一篇

基于非参数相机标定的位移在线检测系统

徐 赢,段发阶,蒋佳佳,许 飞,梁春疆,冯 帆,伯 恩   

  1. (天津大学精密测试技术及仪器国家重点实验室,天津300072)
  • 收稿日期:2014-06-13 出版日期:2015-06-15 发布日期:2015-06-15
  • 作者简介:徐 赢(1990 - ),女,硕士研究生,主研方向:计算机视觉;段发阶,教授、博士生导师;蒋佳佳,讲师;许 飞、梁春疆、冯 帆、 伯 恩,硕士研究生。
  • 基金项目:

    国家“863”计划基金资助项目(2013AA102402);国家自然科学基金资助项目(51275349)。

Displacement Online Detection System Based on Nonparametric Camera Calibration

XU Ying,DUAN Fajie,JIANG Jiajia,XU Fei,LIANG Chunjiang,FENG Fan,BO En   

  1. (State Key Laboratory of Precision Measuring Technology & Instruments,Tianjin University,Tianjin 300072,China)
  • Received:2014-06-13 Online:2015-06-15 Published:2015-06-15

摘要:

物体受外力作用时因发生位移变形而导致结构损坏。针对该问题,以双目视觉原理建立物体三维数学模型,研究相机标定和特征提取,完成特征匹配,恢复标识点三维几何信息,得到物体位移。由于传统的相机标定法在常规尺寸应用中精度较高,一旦超出视场范围其测量精度迅速降低,不适于大视场测量,因此引入一种针对大视场测量的非参数相机标定法,并建立基于非参数的成像模型。通过在不同环境下分别对2 种标定方法进行精度对比实验,结果表明,非参数相机标定法比传统的参数相机标定法精度提高约72. 5% ,且稳定性好,可满足物体位移测量的需求。

关键词: 双目视觉, 特征提取, 特征匹配, 非参数相机标定, 非参数成像, 位移测量

Abstract:

The object under the action of external force tends to have structural damage due to its displacement and deformation. To handle the problem,the three-dimensional mathematic model of pipe must be established with binocular vision principle. By studying camera calibration and feature extraction,matching the feature and restoring 3D geometric information of the marking point,displacement of the object can be obtained. Due to the traditional camera calibration method is high-precision in conventional applications,once beyond the field range,its measurement accuracy decreases rapidly,so it is not suitable for field measurement of the project. This paper introduces a method about nonparametric camera calibration for field measurement,and establishes the imaging model based on the new method. By conducting accuracy comparison experiments according to the calibration method in different environments,and the test results show that,precision of nonparametric camera calibration is approximately 72. 5% higher than the traditional method,as well as good stability,which can meet the measurement requirements of displacement.

Key words: binocular vision, feature extraction, feature matching, nonparametric camera calibration, nonparametric imaging, displacement measurement

中图分类号: